import json import os import subprocess import time import re from pathlib import Path from playwright.sync_api import sync_playwright # --- CONFIG --- QUEUE_FILE = "/app/services/gis_harvester/data/full_queue.json" STORAGE_ROOT = Path("/app/services/gis_harvester/archive") NAS_IP = "192.168.10.105" NAS_PATH = "/volume1/web/legal/gis/raw" BOT_TOKEN = "8725618164:AAH1tGalq-pw1l0t4P5c0sdLkCVJdh7IE7M" CHAT_ID = "197957361" def send_tg_status(msg): subprocess.run(["curl", "-s", "-X", "POST", f"https://api.telegram.org/bot{BOT_TOKEN}/sendMessage", "-d", f"chat_id={CHAT_ID}&text={msg}&parse_mode=markdown"]) def sync_to_nas(local_path, remote_folder): target = f"{NAS_PATH}/{remote_folder}" subprocess.run(["ssh", f"matrixhasyou@{NAS_IP}", f"mkdir -p {target}"]) for f in os.listdir(local_path): src = local_path / f dst = f"{target}/{f}" with open(src, "rb") as fd: subprocess.run(["ssh", f"matrixhasyou@{NAS_IP}", f"cat > '{dst}'"], stdin=fd) def run_harvest(): if not os.path.exists(QUEUE_FILE): print("❌ Очередь не найдена!") return with open(QUEUE_FILE, "r") as f: queue = json.load(f) total = len(queue) print(f"[*] Начало жатвы {total} целей...") with sync_playwright() as p: try: browser = p.chromium.connect_over_cdp("http://127.0.0.1:9222") page = browser.contexts[0].pages[0] for index, item in enumerate(queue): num = item['number'].replace('/', '_').replace(' ', '_')[:50] local_dir = STORAGE_ROOT / num # Заходим в каждое дело try: print(f"[{index+1}/{total}] Обработка {num}...") page.goto("https://my.dom.gosuslugi.ru" + item['href']) page.wait_for_load_state("networkidle") time.sleep(10) # Скриншот local_dir.mkdir(parents=True, exist_ok=True) page.screenshot(path=str(local_dir / "evidence.png"), full_page=True) # Скачивание download_btn = page.get_by_text("Скачать все").first if download_btn.is_visible(): with page.expect_download(timeout=120000) as download_info: download_btn.click() download = download_info.value zip_path = local_dir / "attachments.zip" download.save_as(str(zip_path)) # Синхронизируем sync_to_nas(local_dir, num) status = "✅ ZIP + Скрин" else: sync_to_nas(local_dir, num) status = "⚠️ Только скрин" # Информирование bar = "█" * ((index + 1) * 10 // total) + "░" * (10 - ((index + 1) * 10 // total)) msg = f"🌾 **ЖНЕЦ ГИС ЖКХ**\n\n🎯 {item['number'][:50]}\nСтатус: {status}\n\n[{bar}] {index+1}/{total}" send_tg_status(msg) except Exception as e: print(f"⚠️ Ошибка на {num}: {e}") page.goto("https://my.dom.gosuslugi.ru/citizen-cabinet/#!/appeals/applicant/search") time.sleep(5) browser.close() except Exception as e: send_tg_status(f"🚨 СБОЙ ЖНЕЦА: {str(e)}") if __name__ == "__main__": run_harvest()